Information

Scottish Household Survey 2020: methodology and impact of change in mode

The methodology report for the Scottish Household Survey 2020 telephone survey which discusses the impact of the change in mode.

This document is part of 2 collections


Chapter 5: Impact of the change of mode of approach on respondent profile

The analysis in this chapter looks at the composition of the achieved SHS sample. It compares estimates from the 2019 waves of the SHS to those from 1) the interviews carried out face-to-face in 2020 prior to lockdown 2) the push-to-telephone/video opt-in only sample 3) the push-to-telephone/video telephone number matched sample and 4) the combined push-to-telephone/video sample.

All the analysis is based on weighted data. For the 2020 data, the weights were constructed separately for the three different sample types, independent of each other, as if each were the final achieved sample.

Prior to 2020, calibration weighting was used to derive weights that matched NRS population totals for age bands and sex within each Local Authority. For household weights, this involved all population in responding households and for random adult weights, just those over interviewed. The 2020 weights have attempted to adopt a similar approach, but due to the smaller sample sizes, age groups have been expanded for within local authority calibration targets. The proportion of the population living in each SIMD quintile and 6-fold rural-urban classification were also added as calibration targets in an attempt to address the response rate differences discussed in the last chapter.

The weights for the combined push-to-telephone/video sample were constructed by combining the unweighted opt-in only and telephone number matched samples and then calibrating. They were not constructed by combining the weighted opt-in only and telephone number matched samples. Therefore, it is possible that for some measures the weighted estimate for the combined sample will be higher or lower than the weighted estimates for the opt-in only and telephone number matched samples (e.g. see the estimates of households where the highest income householder is male in Table 5.3)[22].

It is worth emphasising that all surveys are only estimates of what they seek to measure, and the 2018 and 2019 waves will be subject to error and bias. Even so, they are useful as benchmarks to examine changes in the nature of bias affecting SHS estimates.

Overall, 49 key survey measures were included for analysis:

  • 2 Geographic measures: rurality and SIMD quintile.
  • 8 Household measures: tenure, length of time at address, property type, household type, household working status, household income, whether managing financially, and satisfaction with housing.
  • 3 Highest Income Householder measures: Banded age, gender, an economic status.
  • 36 measures from the random adult interview. These are split into 8 where we would expect minimal change between 2019 and 2020 (such as age, gender, and educational attainment and general health), and 28 others that are more likely to have been seen considerable changes to the estimates as a results of external changes, such as changes arising from the Covid-19 pandemic and associated restrictions.

Geographic variables

Table 5.1 compares the 2019 estimates with results from the three different approaches used in 2020 for Rurality and SIMD Quintile[23].

Table 5.1 Rurality and SIMD quintile by wave (weighted, households)
2019 2020 - f2f 2020 – p2tv Opt-in Diff (-2019) 2020 - p2tv Tel Diff (-2019) 2020 - p2tv All Diff (-2019)
Urban/rural indicator
Large Urban 35.1% 35.7% 34.6% -0.5% 35.7% 0.6% 35.2% 0.1%
Other Urban 35.5% 36.3% 36.8% 1.3% 36.0% 0.5% 36.2% 0.7%
Accessible Small Towns 8.7% 8.1% 7.7% -1.0% 9.7% 0.9% 8.7% 0.0%
Remote Small Towns 3.8% 3.6% 2.2% -1.6% 2.5% -1.2% 2.6% -1.2%
Accessible Rural 10.9% 10.4% 11.5% 0.6% 9.8% -1.2% 11.4% 0.5%
Remote Rural 6.0% 5.9% 7.2% 1.2% 6.3% 0.4% 6.0% 0.0%
SIMD Quintile
Most deprived 20.8% 19.9% 20.6% -0.2% 20.4% -0.4% 21.0% 0.2%
2nd 20.6% 21.9% 15.4% -5.2% 18.2% -2.4% 16.7% -3.9%
Middle quintile 20.0% 19.9% 20.0% 0.0% 13.7% -6.3% 18.7% -1.3%
4th 19.5% 19.9% 25.9% 6.3% 29.0% 9.5% 25.1% 5.6%
Least deprived 19.1% 18.5% 18.2% -0.9% 18.7% -0.4% 18.5% -0.6%
N 10,577 1,545 1,718 1,313 3,031

In summary:

  • Urban/Rural indicator – the estimates from the push-to-telephone/video opt-in sample and the telephone matched sample are similar to the 2019 results. The maximum difference is 1.6 percentage points and there is no obvious pattern to the differences across the categories.
  • SIMD Quintile – although SIMD is used in the weighting approach, there are considerable differences between the 2019 estimates and the push-to-telephone/video opt-in sample. Because of the small sample size, calibration targets were only set for quintile 1, quintiles 2 to 4 combined, and quintile 5.

There should be very little change in either of these measures year on year. SIMD, rurality and local authority were all explicitly included in the weighting, and differences are likely to be due to how the calibration weights have been produced[24].

Household measures

Table 5.2 shows the same comparisons for eight household measures.

Table 5.2 Household measures by wave (weighted)
2019 2020 - f2f 2020 – p2tv Opt-in Diff (-2019) 2020 - p2tv Tel Diff (-2019) 2020 - p2tv All Diff (-2019)
Tenure
Owner-occupied 61.6% 61.7% 68.1% 6.5% 79.5% 18.0% 71.8% 10.3%
Social Rented 23.5% 23.8% 15.4% -8.1% 17.4% -6.1% 15.5% -8.0%
Private Rented 13.8% 13.3% 15.2% 1.4% 2.7% -11.0% 11.6% -2.2%
Other 1.1% 1.1% 1.4% 0.3% 0.3% -0.8% 1.0% -0.1%
Length of time at address[25]
Less than a year 11.2% 11.8% 10.9% -0.3% 4.3% -6.8% 8.2% -3.0%
1-3 years 19.6% 19.3% 25.0% 5.4% 16.8% -2.8% 20.8% 1.2%
4-15 years 35.1% 34.4% 36.3% 1.2% 35.5% 0.4% 34.0% -1.0%
Over 15 years 34.1% 34.6% 27.8% -6.4% 43.4% 9.3% 37.0% 2.9%
Property type
House 65.8% 64.1% 62.2% -3.6% 78.2% 12.4% 67.8% 2.0%
Flat 33.6% 35.6% 37.3% 3.7% 21.5% -12.1% 32.1% -1.5%
Other 0.6% 0.3% 0.5% -0.1% 0.2% -0.4% 0.1% -0.5%
Household type
Single adult 20.4% 20.6% 20.3% -0.1% 9.6% -10.8% 16.6% -3.8%
Small adult 20.2% 20.4% 19.3% -0.9% 17.0% -3.2% 18.7% -1.5%
Single parent 4.4% 4.7% 3.5% -0.9% 5.3% 1.0% 3.9% -0.4%
Small family 12.8% 12.4% 14.5% 1.7% 18.1% 5.3% 14.1% 1.3%
Large family 5.2% 5.1% 5.9% 0.8% 8.1% 2.9% 6.9% 1.7%
Large adult 8.8% 9.1% 8.6% -0.2% 15.8% 7.0% 11.9% 3.1%
Older smaller 13.9% 13.1% 14.1% 0.2% 14.4% 0.5% 15.8% 1.9%
Single pensioner 14.4% 14.6% 13.8% -0.6% 11.7% -2.7% 12.1% -2.3%
Household working status
Single working adult 19.6% 20.3% 18.5% -1.1% 17.5% -2.1% 16.5% -3.1%
Non-working single 26.4% 28.6% 24.7% -1.7% 18.2% -8.2% 22.5% -3.9%
Working couple 30.0% 26.9% 31.7% 1.8% 36.6% 6.6% 33.7% 3.8%
Couple, one works 10.5% 11.1% 10.9% 0.5% 15.6% 5.2% 12.4% 1.9%
Couple, neither work 13.6% 13.1% 14.1% 0.5% 12.1% -1.5% 14.9% 1.3%
N 10,577 1,545 1,718 1,313 3,031
Net annual household income
£0-£10,000 8.3% 7.5% 6.9% -1.4% 3.6% -4.7% 6.5% -1.8%
£10,001-£20,000 26.6% 28.0% 22.5% -4.0% 20.0% -6.6% 20.5% -6.1%
£20,001-£30,000 21.3% 21.6% 21.9% 0.7% 22.3% 1.0% 21.6% 0.3%
£30,001-£40,000 15.4% 17.1% 16.9% 1.5% 14.7% -0.7% 16.5% 1.1%
£40,001+ 28.4% 25.8% 31.7% 3.3% 39.4% 10.9% 34.9% 6.5%
Whether struggling financially
Struggling financially 8.8% 7.3% 8.7% -0.1% 2.8% -6.1% 6.9% -2.0%
Satisfaction with housing
Very/fairly satisfied 90.1% 90.0% 92.2% 2.1% 94.6% 4.5% 94.0% 3.8%

This shows the following:

  • Tenure: The overall estimate for owner-occupation among the combined push-to-telephone/video samples is 10 percentage points higher than the 2019 estimate (71.8% compared to 61.6%). Both the push-to-telephone/video samples appear to be considerably biased, with the telephone matched sample showing particularly large differences compared to the 2019 figures: the estimate for owner-occupiers is 18.0 percentage points higher (79.5% compared to 61.6%), social renters are 6.1 percentage points lower (17.4% compared to 23.5%), and the estimate for private renters is 11.0 percentage points lower, dropping from 13.8% to only 2.7%[26]. In the opt-in sample, the estimate for owner-occupation is 6.5 percentage points more than the 2019 estimate (68.1% compared to 61.6%) and social rented is 8.1 lower than the 2019 estimate. However, the estimate for private renters is closer to the 2019 estimate (15.2% compared to 13.8%).
  • Length of time at address: the telephone-matched sample appears very biased against those who have only lived at their address for a short time. In 2019, 11.2% said they had lived at their address for less than a year, and 19.6% had lived there for between 1 and 3 years. The corresponding estimates among the telephone matched sample were only 4.3% for less than a year and 16.8% for between 1 and 3 years. In comparison, the differences between the opt-in sample and the 2019 wave were smaller but still considerable. The revised approach overall reflects the difference between the telephone-matched sample and the 2019 wave, and appears particularly biased against those who have lived at their address for less than one year.
  • Property type: The estimates from the telephone matched sample over-represent those living in houses (78.2% compared to 65.8%) and under-estimate those living in flats (21.5% compared to 33.6%). The estimates from the opt-in only sample are closer to the 2019 wave but with flats over-represented (37.3% compared to 33.6%) and houses under-represented (62.2% compared to 65.8%). Overall, the revised approach over-estimates those living in houses (67.8% compared to 65.8%).
  • Household type: The telephone-matched sample again shows large differences compared to the 2019 wave and appears to over-represent large adult households and small family households and under-represent single adult and small adult households. The differences from the 2019 wave are again smaller for the opt-in sample, with a maximum difference of 1.7 percentage points (for small family households). Overall, single adult and single pensioner households are under-represented in the push-to-telephone/video combined data.
  • Household working status: Reflecting differences in household type, both single working households and non-working single households are under-represented in the push-to-telephone/video combined data. This is seen in both the telephone-matched sample and the opt-in sample, although the differences compared with the 2019 wave are larger for the telephone-matched sample.
  • Household income: Households with a net income of over £40,000 are over-represented in the push-to-telephone/video data, while households in the lower income bands are under-represented. This is much more pronounced in the telephone-matched sample than the opt-in only sample. In the telephone-matched sample, the estimate for households with £40,000+ is 10.9 percentage points more than the 2019 estimate, while the corresponding difference for the opt-in only sample is 3.3%.
  • Whether struggling financially: In the 2019 wave, 8.8% said that they were not managing well financially or were in deep financial trouble. Overall, the estimate from the push-to-telephone/video approach was 6.9%. The estimate from the opt-in sample (8.7%) was much closer to the 2019 figure than the estimate for the telephone matched sample (2.8%). The impact of the pandemic on this measure is likely to have been complex with changes to both income and expenditure patterns, and significant differences across different types of households.
  • Satisfaction with housing: The proportion saying that they are very or fairly satisfied with their housing is 94.0% in the push-to-telephone/video sample compared to 90.1% in 2020.

The opt-in only sample is considerably closer to the 2019 results than the telephone matched-sampled sample on all these measures.

Although there might have been greater than usual change to household formation during the pandemic, we would not expect anything more than minimal changes for these measures. As discussed in Chapter 3 previous face-to-face estimates have matched up well with census data as well as administrative sources in the past. The available administrative data on social housing stock from social landlords and on private rented properties from the Landlord Registration System indicates that the number of households in the social rented and private rented sectors are likely to have remained at relatively similar levels across the 2020 period compared to 2019, with the number of social rented dwellings expected to have seen a slight increase in 2020 due to the increase in recent years in the level of new affordable housing along with the ending of the Right to Buy scheme. Overall, this suggests that the revised push-to-telephone video approach appears to have introduced bias, particularly in relation to tenure and length of time at the address. This would also mean that estimates that are highly correlated with these characteristics are also likely to be affected.

Highest Income Householder measures

Table 5.3 provides comparisons by wave for three variables related to the highest income householder (HIH).

Table 5.3 HIH measures by wave (weighted)
2019 2020 - f2f 2020 – p2tv Opt-in Diff (-2019) 2020 - p2tv Tel Diff (-2019) 2020 - p2tv All Diff (-2019)
HIH Banded age
16-24 4.6% 4.9% 3.7% -0.8% . -4.6% 2.4% -2.1%
25-44 30.3% 30.0% 31.3% 0.9% 31.9% 1.6% 31.0% 0.7%
45-59 28.7% 27.5% 29.2% 0.5% 32.3% 3.6% 29.6% 0.9%
60+ 36.4% 37.6% 35.8% -0.6% 35.8% -0.7% 36.9% 0.5%
HIH Gender
Male 57.7% 57.3% 56.2% -1.4% 57.9% 0.2% 58.0% 0.4%
Female 42.3% 42.7% 43.5% 1.2% 42.1% -0.2% 41.8% -0.5%
HIH Economic status
Self employed 7.5% 8.2% 6.5% -1.0% 6.2% -1.3% 6.3% -1.1%
Employed full time 43.8% 40.0% 44.1% 0.4% 51.7% 8.0% 47.1% 3.3%
Employed part time 7.0% 8.0% 8.7% 1.7% 8.5% 1.4% 7.0% -0.1%
Looking after the home/family 1.8% 1.4% 0.8% -1.1% 0.5% -1.3% 0.9% -0.9%
Retired from work 28.3% 28.9% 28.8% 0.4% 28.1% -0.2% 29.3% 1.0%
Unemployed 2.5% 2.5% 4.2% 1.7% 1.0% -1.5% 3.2% 0.7%
In further/higher education 2.6% 4.1% 2.5% -0.1% 0.0% -2.5% 1.7% -0.9%
Permanently sick or disabled 5.0% 5.4% 3.5% -1.5% 3.9% -1.1% 3.6% -1.4%
Short-term illness or injury 1.0% 0.6% 0.6% -0.4% 0.0% -0.9% 0.6% -0.4%

As with the household factors, the differences between the estimates from the telephone-matched sample and the 2019 wave tend to be larger than between the opt-in sample and the 2019 wave. In summary:

  • HiH age: Overall, younger HIHs are under-represented in the push-to-telephone/video data. This pattern is very marked among the telephone matched sample. This is likely to be because young people are less likely to have landlines and therefore to be in the telephone-matched sample. Among the opt-in sample, the differences are much smaller[27].
  • HiH Gender: The push-to-telephone/video data is very similar to the 2019 figures. This is not surprising given the weighting strategy. Unusually, the opt-in sample estimate was further from the 2019 figures than the estimate from the telephone matched sample.
  • HiH Economic Status: The push-to-telephone/video data over-represents those employed full-time by 3.3 percentage points. The telephone-matched sample is 8.0 percentage points above the 2019 estimate. In contrast, the corresponding difference for the opt-in only sample is 0.4 percentage points.

The change in approach appears to have introduced bias in terms of the age profile of Highest Income Householders, with younger age groups under-represented. The level of change in the estimate for gender is smaller, and economic activity may have been impacted by the pandemic. However, again the opt-in sample appears to be closer to the 2019 estimates than the telephone-matched sample.

Random Adult measures

Table 5.4 shows how the estimates from the revised approach compare to 2019 for eight random adult measures that we would expect to be relatively stable between the waves – age, gender, ethnicity, attainment, general health, disability, access to greenspace and personal use of the internet:

  • Banded age: Age is used in the weighting, and the differences are relatively small with no clear pattern.
  • Gender: Gender is also used in the weighting[28]. There was very little difference (0.4 percentage points) in the estimates for men and women between the 2020 push-to-telephone/video data and the 2019 wave.
  • Ethnicity: Compared to the 2019 figures, the telephone-matched sample underestimates minority ethnic groups[29] (0.1% compared to 4.2%). The estimate for minority ethnic groups from opt-in only sample is much closer to the 2019 figure (5.1% compared to 4.2%).
  • Attainment: Overall, the push-to-telephone/video approach under-represents those who have no qualifications (10.9% compared to 15.3%) and those with the lowest attainment level (13.7% compared to 17.4%) and over-represents those with degrees or professional qualifications (40.0% compared to 32.0%). Unlike most other measures, the estimates from the opt-in sample are further from the 2019 estimates than those from the telephone matched sample and appear particularly biased towards those with degrees or professional qualifications (44.1% compared to 32.0%). This is likely to be because those who are more highly educated may be more familiar with surveys and more interested in taking part without needing further convincing from an interviewer either at the doorstep or on the telephone.
  • General health: In 2019, 8.4% described their general health as bad or very bad. The corresponding figure for the push-to-telephone/video approach overall was 5.6%, with the estimate among the telephone matched sample slightly further from the 2019 estimates than the opt-in sample (5.5% and 6.4% respectively).
  • Disability: Overall, the push-to-telephone/video approach produced a slightly lower estimate of disabled adults than the 2019 estimate (22.9% compared to 24.4%)
  • Proximity to greenspace: The push-to-telephone/video approach produced a slightly higher estimate of being within 5 minutes of greenspace than the 2019 estimate (67.8% compared to 65.5%). It is possible that perceptions of proximity to greenspace may have changed over lockdown.
  • Personal use of internet: The push-to-telephone/video approach also gave a higher estimate of using the internet for personal use (91.5% compared to 87.4%) with the estimate for the opt-in only sample (93.1%) higher than the estimate from the telephone-matched sample (89.7%). This measure may have been less stable than the other measures detailed above. The estimates may be reflecting a real change in internet use over lockdown. However, it is also likely to reflect that the easiest way to opt-in to the survey was online.

These comparisons again show that the revised approach appears to have introduced bias, and that the telephone matched sample estimates tend to be further from the 2019 estimates than those from the opt-in only sample. A key exception is educational attainment, where the opt-in only sample appears more biased.

Table 5.4 Random adult measures by wave (weighted)
2019 2020 - f2f 2020 – p2tv Opt-in Diff (-2019) 2020 - p2tv Tel Diff (-2019) 2020 - p2tv All Diff (-2019)
Banded age
16-24 11.1% 12.5% 12.5% 1.4% 12.5% 1.4% 12.5% 1.4%
25-44 33.0% 31.5% 31.5% -1.5% 31.5% -1.5% 31.5% -1.5%
45-59 25.6% 24.2% 25.2% -0.4% 23.7% -2.0% 24.0% -1.6%
60+ 30.3% 31.9% 30.9% 0.5% 32.4% 2.1% 32.1% 1.7%
Gender
Man/Boy 48.1% 48.0% 48.6% 0.5% 48.3% 0.1% 48.5% 0.4%
Woman/Girl 51.8% 51.9% 51.2% -0.6% 51.7% -0.1% 51.4% -0.4%
Ethnicity
White Scottish/British 88.9% 87.6% 89.1% 0.2% 94.8% 6.0% 90.3% 1.4%
White other[30] 6.8% 8.1% 5.8% -1.1% 5.0% -1.8% 5.3% -1.5%
Minority ethnic groups[31] 4.2% 4.2% 5.1% 0.9% 0.1% -4.1% 4.3% 0.1%
Highest educational attainment
None 15.3% 15.0% 10.0% -5.3% 13.2% -2.1% 10.9% -4.4%
Level 1 - O grade etc 17.4% 16.6% 11.7% -5.7% 19.1% 1.7% 13.7% -3.7%
Level 2 - Higher, A 16.7% 15.0% 18.8% 2.0% 18.7% 2.0% 18.8% 2.1%
Level 3 - HNC/HND 12.8% 15.0% 12.6% -0.2% 13.4% 0.6% 12.8% 0.0%
Degree or prof qual 32.0% 33.5% 44.1% 12.1% 30.2% -1.8% 40.0% 8.0%
Other qualification 5.0% 3.5% 2.2% -2.8% 4.4% -0.6% 3.2% -1.9%
General health
General health bad or very bad 8.4% 6.5% 6.4% -2.0% 5.5% -2.9% 5.6% -2.8%
Disability
Disabled 24.4% 24.4% 23.4% -1.0% 25.5% 1.2% 22.9% -1.5%
Non-disabled 75.2% 75.1% 76.5% 1.2% 73.1% -2.1% 76.8% 1.6%
Greenspace
Within 5 mins of greenspace 65.5% 66.1% 70.4% 4.9% 63.5% -2.0% 67.8% 2.3%
Personal use of internet
Used internet for personal use 87.4% 89.6% 93.1% 5.7% 89.7% 2.2% 91.5% 4.1%

Table 5.5 shows results for a variety of other key measures, such as satisfaction with services, cultural attendance, and other attitudinal and behavioural measures that are more likely to have changed during lockdown. Differences on most of these variables are likely to be influenced by the differences between modes in the profile of the respondents. However, they are also more likely to have changed during the pandemic than other estimates. Therefore, they are less illuminating in relation to the impact of the change of approach, since it is even more challenging to assess how much of any observed change is attributable to the change in approach rather than to external circumstances.

In terms of the attitudinal measures, the 2020 estimates compared with the 2019 figures suggest an improvement in relation to satisfaction with services and ratings of local neighbourhood as a good place to live. Similarly, community cohesion measures, such as being able to rely on neighbours, suggest an improvement. The cultural attendance measure shows a considerable decrease, as would be expected because of the pandemic and associated restrictions. Conversely, visits to the outdoors show a marked increase, which again would be expected given the reduction in indoor leisure opportunities and emphasis on meeting outdoors rather than indoors, when restrictions permitted.

The number of people reporting feeling lonely increased considerably.

Table 5.5 Additional random adult measures by wave (weighted)
2019* 2020 - f2f 2020 – p2tv Opt-in Diff (-2019) 2020 - p2tv Tel Diff (-2019) 2020 - p2tv All Diff (-2019)
Culture and Heritage
Cultural attendance 81.0% 80.8% 48.5% -32.5% 35.4% -45.6% 44.2% -36.8%
Cultural participation 75.1% 76.1% 84.0% 8.9% 82.0% 6.9% 83.2% 8.1%
Cultural engagement 90.3% 89.2% 86.8% -3.5% 85.2% -5.1% 86.4% -3.8%
Physical Activity and Sport
Participated in sport in last 4 weeks 79.6% 81.2% 86.9% 7.3% 82.8% 3.2% 85.9% 6.3%
Discrimination and Harassment
Experienced either discrimination or harassment 9.3% 10.1% 9.2% -0.1% 6.8% -2.5% 8.3% -1.0%
Satisfaction with local services
Satisfied with local health services (excluding no opinion) 79.7% 78.3% 88.4% 8.7% 87.5% 7.8% 88.3% 8.6%
Satisfied with local schools (excluding no opinion) 73.2% 73.1% 76.4% 3.2% 83.4% 10.2% 78.2% 5.0%
Satisfied with public transport (excluding no opinion) 67.8% 63.6% 68.5% 0.7% 73.2% 5.3% 69.9% 2.0%
Satisfied with all three services (no opinion for up to two) 52.6% 48.7% 59.3% 6.7% 64.4% 11.8% 60.5% 8.0%
Outdoors
One+ visits to the outdoors 56.0% 58.3% 79.8% 23.8% 76.9% 20.8% 78.8% 22.8%
Social capital
Feels lonely some, most, almost all or all of the time[32] 21.3% 22.0% 36.2% 14.9% 34.4% 13.1% 34.7% 13.4%
Meets socially at least once a week[33] 72.6% 68.8% 43.3% -29.3% 41.5% -31.1% 42.9% -29.7%
Volunteering
Volunteered 26.0% 31.5% 27.2% 1.2% 24.0% -2.0% 25.6% -0.4%
Provided unpaid help to improve their local environment[34] 4.5% 4.4% 7.8% 3.3% 7.9% 3.4% 8.0% 3.5%
Rating of neighbourhood
Rating of neighbourhood as very good 57.0% 53.4% 59.6% 2.7% 59.6% 2.7% 59.1% 2.2%
Rating of neighbourhood as fairly good 37.2% 41.1% 36.1% -1.1% 37.7% 0.5% 37.2% 0.0%
Community belonging
Very/fairly strong feeling on belonging to immediate neighbourhood 77.8% 75.9% 78.3% 0.5% 87.0% 9.2% 80.9% 3.1%
Agreement with statements about local neighbourhood
If I was alone and needed help, I could rely on someone in this neighbourhood to help me 85.4% 84.0% 85.8% 0.4% 90.0% 4.6% 87.5% 2.1%
If my home was empty, I could count on someone in this neighbourhood to keep an eye on my home 84.8% 83.4% 85.7% 0.9% 92.7% 7.9% 87.6% 2.8%
I feel I could turn to someone in this neighbourhood for advice or support 78.4% 74.9% 78.0% -0.4% 83.1% 4.7% 80.0% 1.6%
In an emergency, I would offer to help people in my neighbourhood who might not be able to cope well 89.7% 89.4% 90.9% 1.2% 92.8% 3.0% 92.1% 2.4%
This is a neighbourhood where people are kind to each other 82.8% 81.7% 87.2% 4.5% 92.6% 9.9% 89.1% 6.3%
This is a neighbourhood where most people can be trusted 78.6% 76.9% 83.2% 4.6% 87.2% 8.5% 84.0% 5.4%
There are welcoming places and opportunities to meet new people 51.7% 49.1% 54.5% 2.9% 49.9% -1.8% 53.9% 2.2%
There are places where people can meet up and socialize 57.1% 55.9% 61.9% 4.8% 58.1% 1.0% 61.1% 4.0%
This is a neighbourhood where people from different backgrounds get on well together 69.2% 69.1% 75.6% 6.4% 75.1% 5.9% 76.7% 7.5%
This is a neighbourhood where local people take action to help improve the neighbourhood 57.3% 58.3% 65.8% 8.5% 66.4% 9.1% 67.1% 9.7%
I can influence decisions affecting my local area 17.8% 18.1% 23.2% 5.4% 23.5% 5.7% 24.5% 6.7%

Summary

Overall, for most variables that we would expect to be relatively stable, the differences between the 2019 wave and the 2020 push-to-telephone/video approach, after corrective weighting, were relatively small. However, for a selection of key variables, the observed changes in estimates would not be expected. These include tenure, length of time at property, and educational attainment.

In general, the estimates from the telephone matched sample are substantially further from the 2019 figures than those from the opt-in sample, with under-representation of younger highest income householders, those in social rented and private rented housing, and those who have lived in their current address for a short period of time.

Note that this is despite the response rate for the opt-in sample being considerably lower than the telephone matched sample and is a reminder that higher response rates do not necessarily lead to lower non-response bias.

There is one notable exception. The opt-in only sample appears further from the 2019 estimates on educational attainment, with those with degree level qualifications over-represented.

Contact

Email: shs@gov.scot

Back to top